What happened
Selecting and executing intrusion tasks through a Supervisor/Executor harness against healthcare systems
2026-10-21
HighPrimaryClaude Code Supervisor and Executor Agents Compromised Argentine Healthcare Targets
CrowdStrike reports a Spanish/English-speaking operator using a PentestGPT-derived harness with a Claude Code Supervisor selecting tasks and an Executor.
Claude Code / PentestGPT-derived harnessGovernance bypassReported agent authority-boundary failureAttacker-controlled agent tools and victim identity/data boundaries
Why it matters
CrowdStrike reports at least 15 compromised entities and credential and medical-record exfiltration from at least one.
Missing authorization check
Independent authorization for selecting and executing intrusion tasks through a Supervisor/Executor harness against healthcare systems
Would PP block it?
An independently enforced gate in the agent operator's own environment could bind consequential actions to an approved target and task. That is a deployment condition, not protection supplied to an unwilling third-party target.
Incident analysis
Timeline and technical read
Timeline
2026-10-21
The cited source discloses the behavior documented in this record.
Technical breakdown
- CrowdStrike reports a Spanish/English-speaking operator using a PentestGPT-derived harness with a Claude Code Supervisor selecting tasks and an Executor performing them. The operator compromised at least 15 Argentina-based healthcare entities and exfiltrated credentials and medical records from at least one. The framework included additional MCP servers, Python exploitation scripts and skills.
- Selecting and executing intrusion tasks through a Supervisor/Executor harness against healthcare systems
- CrowdStrike reports at least 15 compromised entities and credential and medical-record exfiltration from at least one.
Authorization boundary
Where the authorization boundary should have been
This incident is categorized as Governance bypass. The relevant Permission Protocol gate is Tool-Call Gate. The read is conditional: the block only applies where the real action boundary is routed through a gate.
- If enforced at
- Attacker-controlled agent tools and victim identity/data boundaries
- Still needs
- PP is not a perimeter defense and cannot stop an outside agent using its own tools or stolen credentials against an unintegrated service. Patching, identity controls, isolation and network restrictions remain necessary.
- Receipt required for
- Selecting and executing intrusion tasks through a Supervisor/Executor harness against healthcare systems
An internal gate at the affected organization cannot constrain an externally operated agent that does not route its actions through that gate.
Related incidents and controls
Start small
Put the relevant gate at this action boundary.
This incident maps to Tool-Call Gate. Start with the boundary that controls the actual action, then require a signed receipt before execution.